Road greening carbon sink dynamic evaluation system and method based on IPCC

Abstract

The invention discloses a road greening carbon sink dynamic evaluation system and method based on IPCC, and the system comprises a data collection layer which is used for obtaining the multi-source heterogeneous data of a road area through a space-air-ground integrated monitoring network; the data storage and processing layer is connected with the data acquisition layer and is used for carrying out preprocessing, space-time registration and standardization on the multi-source heterogeneous data and storing the multi-source heterogeneous data to a forest database and a management decision database; the model calculation layer is connected with the data storage and processing layer and used for executing carbon sink dynamic evaluation and prediction based on the preprocessed data; and the application service layer is connected with the model calculation layer and is used for providing carbon sink visualization, carbon asset management and decision support services for users. According to the invention, accurate and dynamic evaluation of the vegetation carbon sequestration of different road sections along the highway is realized.

Description

Technical Field

[0001] This invention relates to the field of climate change mitigation technology, and in particular to a dynamic assessment system and method for carbon sinks in highway greening based on the IPCC. Background Technology

[0002] Global climate change has become a major challenge facing humanity, and the transportation sector is one of the main sources of greenhouse gas emissions. Statistics show that the transportation sector's carbon emissions are second only to the energy sector globally, and are also at a high level in my country. To address this challenge, highway greening, as an important carbon sink method, is gradually gaining attention. However, existing carbon sink assessment technologies for highway greening have many shortcomings: Low data acquisition accuracy: Traditional methods mainly rely on manual surveys and periodic measurements, which are not only time-consuming and labor-intensive, but also difficult to obtain continuous and comprehensive data. For example, existing highway carbon sink assessments are often based on limited sample points and static data, failing to reflect the dynamic processes and spatial heterogeneity of vegetation growth. Although remote sensing technology has been applied, it is mostly limited to a single data source (such as optical remote sensing), easily affected by weather conditions, and lacks ground validation data, resulting in significant uncertainty in the assessment results.

[0003] Poor model adaptability: Existing carbon sink assessment models mostly employ empirical coefficient methods or static biomass equations, failing to fully consider the growth characteristics of different tree species, tree age structure, and the influence of environmental factors (such as soil moisture, nutrient status, and meteorological conditions). For example, some models simply establish a linear relationship between carbon storage and tree diameter at breast height (DBH) and tree height, ignoring the nonlinear process of vegetation growth and the dynamic impact of climate change. Although the IPCC (Intergovernmental Panel on Climate Change) provides a basic framework and default values ​​for carbon sink calculations, it is designed for national or regional scale applications. Directly applying it to carbon sink assessments of linear engineering projects like highways would introduce significant errors.

[0004] Insufficient dynamic monitoring capabilities: Existing systems often lack real-time monitoring and dynamic updating capabilities. Carbon sink assessment results typically represent a static situation at a specific point in time and cannot reflect the dynamic impacts of seasonal changes, extreme weather events (such as droughts and torrential rains), and human activities (such as pruning and replanting) on ​​carbon sink capacity. Furthermore, existing technologies struggle to predict future carbon sink potential, limiting their application in long-term carbon asset management.

[0005] Credibility and transparency issues: Traditional carbon sink assessment methods suffer from problems such as easy data tampering, opaque processes, and difficulty in verifying results, making it difficult to meet the high standards of data reliability and transparency required by the carbon trading market. This makes it difficult for many highway greening carbon sink projects to obtain international carbon credit certifications (such as VCS and GS standards), hindering their entry into the carbon trading market.

[0006] Low system integration: Current carbon sequestration assessments are mostly limited to a single link or single element, lacking comprehensive consideration of the entire life cycle of a highway (pre-construction, construction, operation, maintenance, and demolition). For example, there are very few carbon sequestration assessment systems designed specifically for highways that can simultaneously take into account factors such as roadside vegetation, wetland soil, and carbon emission offsetting from vehicles traveling on the highway.

[0007] In the field of carbon sequestration assessment for highway greening, the IPCC standard method, process model simulation method, and remote sensing inversion method are three mainstream technical approaches, but they all have significant limitations in terms of adaptability and accuracy to highway scenarios. The IPCC standard method is a carbon sink estimation method developed based on the National Greenhouse Gas Inventory Guidelines published by the Intergovernmental Panel on Climate Change (IPCC). Its core logic is to use default static carbon sequestration factors at the national or regional level (such as annual carbon sequestration per unit area of ​​forest and carbon conversion coefficient of herbaceous plant biomass), combined with basic data such as vegetation area and stand age obtained from manual plot surveys, to calculate the total carbon sink using the multiplication formula "carbon sequestration factor × vegetation area". The main limitation of this method lies in its lack of adaptability to different scenarios. Highway greening is characterized by "linear fragmentation." The soil in areas such as slopes and median strips is infertile and sunlight is uneven. The carbon sequestration capacity of vegetation is significantly lower than that of natural ecosystems. However, the IPCC standard method directly applies regional average parameters, resulting in generally overestimated estimates. The carbon absorption factor has a long update cycle of 5-10 years, which cannot reflect the short-term fluctuations in the carbon sequestration capacity of vegetation along the highway due to maintenance measures (such as pruning and replanting) or pests and diseases. Manual sample plot surveys rely on periodic on-site sampling. For long-distance highways that cross regions, the cost of setting up sample plots is high and the representativeness is insufficient, further amplifying the error.

[0008] Process modeling simulation methods, represented by BEPS (Boreal Ecosystem Productivity Simulator) and In TEC (Integrated Terrestrial Ecosystem Carbon Cycle Model), are assessment methods that use mathematical equations to simulate physiological processes such as vegetation photosynthesis, respiration, and soil carbon cycling. The technical approach involves inputting parameters such as meteorological data (temperature, precipitation), soil data (organic matter content, pH), and vegetation physiological data (leaf area index, photosynthetically active radiation utilization rate), and calculating carbon sinks by simulating the carbon exchange process between plants, soil, and the atmosphere. The core limitation of this method lies in its insufficient dynamic response capability: the model relies on fixed static parameter inputs, such as preset annual average temperature and soil moisture content, and cannot respond in real time to carbon sink fluctuations caused by local microclimates (such as strong winds at tunnel vents) or short-term extreme weather (such as heavy rain and high temperatures) along the highway; common human interventions in highway greening (such as regular irrigation and fertilization) are not included in the model parameter system, resulting in a deviation of more than 30% between the simulation results and the actual carbon sink; the model calculation requires massive basic data support, but it is difficult to obtain detailed meteorological and soil data along the highway, and regional substitute data are often used in practical applications, further reducing accuracy.

[0009] Remote sensing inversion is a technique that uses satellite imagery such as the Landsat series (30-meter spatial resolution) and Sentinel-2 (10-meter spatial resolution) to indirectly estimate carbon sinks by extracting vegetation indices (such as NDVI and EVI). Its principle is based on the statistical correlation between vegetation indices and biomass, establishing empirical formulas (such as "carbon sink = 0.8 × NDVI + 0.05") to convert the spectral information of remote sensing imagery into carbon sink data. For example, by calculating the NDVI value along a highway using Sentinel-2 imagery and substituting it into a preset formula, monthly or quarterly carbon sinks can be obtained. The limitations of this method lie in the contradiction between spatial resolution and temporal updates: the width of roadside green belts is mostly 3-5 meters, while the spatial resolution of mainstream satellite imagery at 30 meters cannot distinguish between roadside vegetation and surrounding land use types, easily leading to the problem of "mixed pixels" and causing an overestimation of carbon sequestration in edge areas; the satellite revisit period (16 days for Landsat and 5 days for Sentinel-2) and cloud cover effects often extend the actual effective data update cycle to quarters or years, making it impossible to capture rapid carbon sequestration changes during peak vegetation growth seasons (such as summer); the empirical formula is highly regional, and the "vegetation index-carbon sequestration" relationship established in a certain area has poor applicability to road sections with different soil types and vegetation compositions, requiring recalibration and lacking universality. Summary of the Invention

[0010] The purpose of this invention is to overcome the problems of strong staticity, insufficient handling of regional differences, weak data real-time performance, and poor scenario adaptability in existing highway greening carbon sink assessment technologies. It proposes a dynamic assessment system and method for highway greening carbon sinks based on IPCC to achieve accurate, dynamic, and transparent assessment of highway greening carbon sink capacity.

[0011] On the one hand, to achieve the above objectives, this invention provides a dynamic assessment system for carbon sequestration in highway greening based on the IPCC, comprising: The data acquisition layer is used to acquire multi-source heterogeneous data of the highway area through an integrated air-space-ground monitoring network. The data storage and processing layer, connected to the data acquisition layer, is used to preprocess, spatiotemporally register and standardize the multi-source heterogeneous data, and store it in the forest database and management decision database. The model calculation layer, connected to the data storage and processing layer, is used to perform dynamic assessment and prediction of carbon sequestration based on the preprocessed data. The application service layer, connected to the model calculation layer, is used to provide users with carbon sink visualization, carbon asset management, and decision support services.

[0012] Preferably, the data acquisition layer includes: The satellite remote sensing module is used to automatically interpret tree species classification, crown area and vegetation index of roadside vegetation over a large area by using high-resolution multispectral and radar satellite imagery and combining deep learning algorithms. The UAV remote sensing module is used to perform detailed scanning of key road sections by UAVs equipped with hyperspectral sensors and lidar to obtain high-precision information on tree height and canopy structure. A ground-based sensor network, deployed along highways, is used to monitor soil organic carbon content, soil moisture, temperature, and atmospheric CO2 concentration in real time. Manual survey terminals are used to assist in data collection and verification.

[0013] Preferably, the model calculation layer includes: The IPCC carbon sequestration module is used to calculate the carbon storage and annual carbon sequestration of vegetation and soil based on the basic methods of IPCC. The multimodal data fusion module is used to fuse modal information from different data sources to generate a unified feature vector; The dynamic assessment and prediction module, built on a machine learning model, is used to predict future carbon sink trends and carbon sink benefits under different management strategies using historical data.

[0014] Preferably, the multimodal data fusion module uses a deep learning model based on an attention mechanism to generate a unified feature vector, specifically: F_fused = Σ (α_i F_i); In the formula, F_i is the feature vector of the i-th mode, α_i is the weight learned by the model, and F_fused is the feature vector.

[0015] Preferably, the machine learning model used in the dynamic evaluation and prediction module is a multilayer perceptron, and the construction of the dynamic evaluation and prediction module includes: The input layer receives feature vectors including tree species, tree height, diameter at breast height, soil parameters, and meteorological data; The transformation is performed through at least one hidden layer containing a non-linear activation function; The output layer outputs the predicted value of future carbon sink benefits.

[0016] Preferably, the application service layer includes a blockchain evidence storage module, which is used to immutably store key monitoring data, model calculation processes and evaluation results through smart contracts, and generate digital credentials containing geofences and timestamps.

[0017] On the other hand, to achieve the above objectives, the present invention also provides a dynamic assessment method for carbon sequestration in highway greening based on the IPCC, comprising: Multi-source data on roadside vegetation and wetlands are collected through an integrated air-space-ground monitoring network, and then preprocessed and standardized. Based on the IPCC methodology and preprocessed data, the vegetation carbon storage, soil carbon storage, and annual carbon sequestration of each area along the highway at the current time point are calculated. Multimodal data is input into a trained machine learning prediction model to dynamically predict future carbon sink trends and simulate carbon sink gains under different management strategies. Input data, model parameters, calculation process and evaluation results are stored on the blockchain to generate carbon sink assessment reports or digital certificates; The parameters of the machine learning prediction model are dynamically optimized based on continuously monitored new data to form a closed-loop optimization system.

[0018] Preferably, the vegetation carbon storage is calculated as follows: C_vegetation = A × B × CF; In the formula, A is the timber volume in m³; B is the timber density in t / m³; CF is the carbon content coefficient, which is dimensionless; and C_vegetation is the vegetation carbon storage.

[0019] Preferably, the training data for the machine learning prediction model includes historical forest growth data, historical environmental monitoring data, and historical management decision data; the management strategies include changing tree species, adjusting irrigation plans, or strengthening wetland maintenance.

[0020] Preferably, the digital credential is a non-fungible token, which embeds geofence information, timestamps, and confidence levels corresponding to the evaluation results.

[0021] Compared with the prior art, the present invention has the following advantages and technical effects: (1) Significantly improved assessment accuracy: Through multi-source data fusion and machine learning model optimization, the accuracy of carbon sink assessment has been greatly improved. For example, the system can identify tree species with strong carbon sequestration capacity and conduct key monitoring and assessment on them; compared with traditional methods, the uncertainty of its assessment results has been reduced by more than 30%, and it can better reflect the true carbon sink capacity of highway greening; (2) It realizes dynamic assessment and prediction of carbon sequestration capacity: The system changes the shortcomings of traditional static assessment, and can present the changes in highway carbon sequestration on a daily, monthly and yearly basis, and can predict future trends. For example, the system can predict the carbon sequestration gain in the next 5 years after replanting high carbon sequestration tree species, or assess the negative impact of extreme drought events on current carbon sequestration capacity, providing the possibility for adaptive management; (3) Improved assessment efficiency and automation level: The integrated air-space-ground automatic monitoring network has replaced more than 80% of the manual survey work, making large-scale and high-frequency carbon sink monitoring possible. The data collection, processing, calculation and visualization have basically achieved full-process automation, shortening the assessment cycle from the traditional "year" to "week" or even "day", and significantly improving efficiency. (4) Enhanced reliability and tradability of assessment results: The introduction of blockchain technology ensures the immutability and traceability of all key data, giving the assessment results the credibility required by the carbon trading market. Each carbon sink assessment report generated by the system carries a digital fingerprint, laying a solid technical foundation for its transformation into tradable carbon assets. This will activate the asset attributes of highway greening carbon sinks, bringing additional economic benefits to highway operators and forming a virtuous cycle. (5) It provides scientific decision support and optimizes carbon sink management: The system's predictive function allows managers to conduct "if-then" scenario analysis. For example, it can compare the carbon sink benefits and costs of "replanting oleander" and "replanting forsythia" over the next ten years, thereby selecting the optimal management decision to maximize carbon sink benefits. In addition, the system can also identify road sections with weak carbon sink capacity and guide precise ecological restoration and improvement. Attached Figure Description

[0022] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of a dynamic assessment system for carbon sinks in highway greening based on IPCC, according to an embodiment of the present invention. Figure 2 This is a flowchart of a dynamic assessment method for carbon sinks in highway greening based on IPCC, according to an embodiment of the present invention. Figure 3 This is a flowchart illustrating the workflow of the multimodal data fusion module in an embodiment of the present invention. Figure 4 This is a schematic diagram of the blockchain evidence storage process according to an embodiment of the present invention. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0025] This embodiment proposes a dynamic assessment system for carbon sequestration in highway greening based on the IPCC, such as... Figure 1 ,include: The data acquisition layer is used to acquire multi-source heterogeneous data of the highway area through an integrated air-space-ground monitoring network. The data storage and processing layer, connected to the data acquisition layer, is used to preprocess, spatiotemporally register and standardize the multi-source heterogeneous data, and store it in the forest database and management decision database. The model calculation layer, connected to the data storage and processing layer, is used to perform dynamic assessment and prediction of carbon sequestration based on the preprocessed data. The application service layer, connected to the model calculation layer, is used to provide users with carbon sink visualization, carbon asset management, and decision support services.

[0026] Specifically, this embodiment integrates IPCC standardized methodologies with dynamic monitoring technologies to construct an assessment system that can accurately adapt to the linear fragmentation scenario of highways. It addresses the problems of estimation bias and delayed carbon absorption factor updates caused by directly applying regional average parameters in the IPCC standard method; it overcomes the shortcomings of process model simulation methods, which cannot respond in real time to local microclimates and human interventions and rely on static parameter inputs; and it overcomes the limitations of remote sensing inversion methods, which struggle to capture the precise carbon sink dynamics of highway greenbelts due to the contradiction between spatial resolution and time update cycles. Ultimately, it achieves dynamic and accurate assessment of vegetation carbon sinks in different sections along highways (such as roadbed slopes, tunnel slopes, and median strips), providing reliable metrological basis for the scientific management of highway greenbelt carbon sinks, optimization of maintenance strategies, and inclusion in carbon market trading, thereby improving the timeliness, accuracy, and applicability of highway ecosystem carbon sink assessments.

[0027] Furthermore, the data acquisition layer includes: The satellite remote sensing module is used to automatically interpret tree species classification, crown area and vegetation index of roadside vegetation over a large area by using high-resolution multispectral and radar satellite imagery and combining deep learning algorithms. The UAV remote sensing module is used to automatically interpret information such as tree species classification, crown area, and vegetation index (such as NDVI) of vegetation over a large area of ​​roadside using high-resolution multispectral and radar satellite imagery and deep learning algorithms. A ground-based sensor network, deployed along highways, is used to monitor soil organic carbon content, soil moisture, temperature, and atmospheric CO2 concentration in real time. Manual survey terminals are used to assist in the collection and verification of data, such as field measurements of tree diameter at breast height (DBH).

[0028] Furthermore, the data storage and processing layer utilizes a cloud platform for the storage and management of massive amounts of data. The core of this layer consists of a forest database and a management decision database. Data preprocessing is responsible for spatiotemporal registration, missing value imputation, outlier handling, and data standardization of multi-source data, forming a time-series dataset with a unified format.

[0029] Furthermore, the model computation layer includes: The IPCC carbon sequestration module is used to calculate the carbon storage and annual carbon sequestration of vegetation and soil based on the basic methods of IPCC. For example, for vegetation, the calculation path is "biomass = volume × density" and "carbon storage = biomass × carbon content coefficient".

[0030] The multimodal data fusion module is used to fuse modal information from different data sources to generate a unified feature vector, such as... Figure 3 ; Specifically, a cross-modal attention mechanism is adopted to fuse information from different modalities such as optical remote sensing, radar, photosynthetically active radiation, and meteorological data to generate a unified feature vector, which significantly improves the accuracy of carbon storage prediction.

[0031] The dynamic assessment and prediction module, built on a machine learning model, is used to predict future carbon sink trends and carbon sink benefits under different management strategies using historical data. Specifically, a predictive model is built based on machine learning models such as multilayer perceptrons (MLP). This model is trained using historical data (historical forest data, historical management decisions) and can predict changes in carbon sequestration benefits over a future period under different management strategies (such as changing tree species and adjusting irrigation schemes).

[0032] Furthermore, the multimodal data fusion module uses a deep learning model based on an attention mechanism to generate a unified feature vector, specifically: F_fused = Σ (α_i F_i); In the formula, F_i is the feature vector of the i-th mode, α_i is the weight learned by the model, and F_fused is the feature vector.

[0033] The dynamic evaluation and prediction module uses a multilayer perceptron as its machine learning model, and the construction of the dynamic evaluation and prediction module includes: The input layer receives feature vectors including tree species, tree height, diameter at breast height, soil parameters, and meteorological data; The transformation is performed through at least one hidden layer containing a non-linear activation function; The output layer outputs the predicted value of future carbon sink benefits.

[0034] Furthermore, the application service layer includes a blockchain evidence storage module. This module is used to immutably store key monitoring data, model calculation processes, and evaluation results through smart contracts, and generate digital credentials containing geofences and timestamps, such as... Figure 4 .

[0035] Specifically, the application service layer also includes: Carbon Sequestration Visualization Screen: Dynamically displays the spatial distribution and temporal changes of carbon sequestration along the entire highway; Carbon asset management module: Store assessment results on the blockchain, generate credible carbon asset reports, and explore integration with carbon trading platforms; Decision support module: Allows users to set different management scenarios, simulate and predict their carbon sequestration benefits, and recommend the optimal management decisions.

[0036] Organizations using this system primarily include: transportation management departments, environmental protection agencies, highway operating companies, and carbon trading market participants. They can use this system to continuously monitor and quantitatively assess the carbon sequestration effects of highway greening projects, providing data support for low-carbon highway operation and carbon asset development.

[0037] This embodiment also provides a dynamic assessment method for carbon sinks in highway greening based on IPCC, such as Figure 2 ,include: Multi-source data on roadside vegetation and wetlands are collected through an integrated air-space-ground monitoring network, and preprocessed and standardized. The preprocessing includes cleaning, transforming and spatiotemporal registration of the data, and storing it in a unified forest database. Based on the IPCC methodology and preprocessed data, the carbon storage (vegetation carbon pool and soil carbon pool) and annual carbon sequestration of each area along the highway at the current time point are calculated; for example, referring to the study of the Shanghai-Nanjing section of G42 Expressway, the carbon sink of trees, shrubs, herbaceous plants and wetland soils are calculated respectively. Multimodal data is input into a trained machine learning prediction model to predict future trends in carbon sequestration. Simultaneously, different management strategies (such as "replanting goldenrain trees" and "strengthening wetland maintenance") can be set to predict their carbon sequestration gains over a future period, providing a basis for decision-making. Input data, model parameters, calculation process and evaluation results are stored on the blockchain to ensure that the data is tamper-proof and the process is transparent, and finally a carbon sink assessment report or digital certificate (such as NFT) containing geofence, timestamp and confidence level is generated. The parameters of the machine learning prediction model are dynamically optimized based on continuously monitored new data, forming a closed-loop optimization system, so that the evaluation model becomes more and more accurate as data accumulates.

[0038] Furthermore, the carbon storage of the vegetation is calculated as follows: C_vegetation = A × B × CF; In the formula, A is the timber volume in m³; B is the timber density in t / m³; CF is the carbon content coefficient, which is dimensionless; and C_vegetation is the vegetation carbon storage.

[0039] Soil carbon storage is calculated based on measured parameters such as soil organic carbon content and soil bulk density.

[0040] Furthermore, a deep learning model based on an attention mechanism is adopted. The core idea is to perform weighted fusion of data from different modalities, where the attention weights (α_i) reflect the importance of each modal data to the current prediction task. F_fused = Σ (α_i F_i); Where F_i is the feature vector of the i-th mode, and α_i is the weight learned by the model.

[0041] A multilayer perceptron (MLP) network is employed, with its input layer including features (X) such as tree species, tree height, diameter at breast height (DBH), soil parameters, and meteorological data. After nonlinear transformation through multiple hidden layers (H), the final output layer (Y) predicts future carbon sequestration benefits. H1 = σ(W1X + b1); H2 = σ(W2H1 + b2); Y = σ(W_nH_{n-1} + b_n); Where W and b are model parameters, and σ is the activation function.

[0042] Furthermore, the training data for the machine learning prediction model includes historical forest growth data, historical environmental monitoring data, and historical management decision data; the management strategies include changing tree species, adjusting irrigation plans, or strengthening wetland maintenance.

[0043] Furthermore, the digital credential is a non-fungible token, which embeds geofence information, timestamps, and confidence levels corresponding to the evaluation results.

[0044] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A dynamic assessment system for carbon sequestration in highway greening based on IPCC, characterized in that, include: The data acquisition layer is used to acquire multi-source heterogeneous data of the highway area through an integrated air-space-ground monitoring network. The data storage and processing layer, connected to the data acquisition layer, is used to preprocess, spatiotemporally register and standardize the multi-source heterogeneous data, and store it in the forest database and management decision database. The model calculation layer, connected to the data storage and processing layer, is used to perform dynamic assessment and prediction of carbon sequestration based on the preprocessed data. The application service layer, connected to the model calculation layer, is used to provide users with carbon sink visualization, carbon asset management, and decision support services.

2. The IPCC-based dynamic assessment system for carbon sequestration in highway greening according to claim 1, characterized in that, The data acquisition layer includes: The satellite remote sensing module is used to automatically interpret tree species classification, crown area and vegetation index of roadside vegetation over a large area by using high-resolution multispectral and radar satellite imagery and combining deep learning algorithms. The UAV remote sensing module is used to perform detailed scanning of key road sections by UAVs equipped with hyperspectral sensors and lidar to obtain high-precision information on tree height and canopy structure. A ground-based sensor network, deployed along highways, is used to monitor soil organic carbon content, soil moisture, temperature, and atmospheric CO2 concentration in real time. Manual survey terminals are used to assist in data collection and verification.

3. The IPCC-based dynamic assessment system for carbon sequestration in highway greening according to claim 1, characterized in that, The model computation layer includes: The IPCC carbon sequestration module is used to calculate the carbon storage and annual carbon sequestration of vegetation and soil based on the basic methods of IPCC. The multimodal data fusion module is used to fuse modal information from different data sources to generate a unified feature vector; The dynamic assessment and prediction module, built on a machine learning model, is used to predict future carbon sink trends and carbon sink benefits under different management strategies using historical data.

4. The IPCC-based dynamic assessment system for carbon sequestration in highway greening according to claim 3, characterized in that, The multimodal data fusion module uses an attention-based deep learning model to generate a unified feature vector, specifically: F_fused = Σ (α_i F_i); In the formula, F_i is the feature vector of the i-th mode, α_i is the weight learned by the model, and F_fused is the feature vector.

5. The IPCC-based dynamic assessment system for carbon sequestration in highway greening according to claim 3, characterized in that, The dynamic evaluation and prediction module uses a multilayer perceptron as its machine learning model, and the construction of the dynamic evaluation and prediction module includes: The input layer receives feature vectors including tree species, tree height, diameter at breast height, soil parameters, and meteorological data; The transformation is performed through at least one hidden layer containing a non-linear activation function; The output layer outputs the predicted value of future carbon sink benefits.

6. The IPCC-based dynamic assessment system for carbon sequestration in highway greening according to claim 1, characterized in that, The application service layer includes a blockchain evidence storage module, which is used to immutably store key monitoring data, model calculation processes and evaluation results through smart contracts, and generate digital credentials containing geofences and timestamps.

7. A dynamic assessment method for carbon sequestration in highway greening based on IPCC, applied to the system described in any one of claims 1-6, characterized in that, include: Multi-source data on roadside vegetation and wetlands are collected through an integrated air-space-ground monitoring network, and then preprocessed and standardized. Based on the IPCC methodology and preprocessed data, the vegetation carbon storage, soil carbon storage, and annual carbon sequestration of each area along the highway at the current time point are calculated. Multimodal data is input into a trained machine learning prediction model to dynamically predict future carbon sink trends and simulate carbon sink gains under different management strategies. Input data, model parameters, calculation process and evaluation results are stored on the blockchain to generate carbon sink assessment reports or digital certificates; The parameters of the machine learning prediction model are dynamically optimized based on continuously monitored new data to form a closed-loop optimization system.

8. The method according to claim 7, characterized in that, The carbon storage of the vegetation is calculated as follows: C_vegetation = A × B × CF; In the formula, A is the timber volume in m³; B is the timber density in t / m³; CF is the carbon content coefficient, which is dimensionless; and C_vegetation is the vegetation carbon storage.

9. The method according to claim 8, characterized in that, The training data for the machine learning prediction model includes historical forest growth data, historical environmental monitoring data, and historical management decision data; the management strategies include changing tree species, adjusting irrigation plans, or strengthening wetland maintenance.

10. The method according to claim 8, characterized in that, The digital credential is a non-fungible token, which embeds geofence information, timestamps, and confidence levels corresponding to the evaluation results.

Images